A new withanolide from the roots of Withania somnifera - NOPR

0 downloads 0 Views 160KB Size Report
Quantitative structure-activity relationships (QSAR) study on a series of (substituted 1, 2-dihydro)4–thiazolidinones and 2-azetidinones bearing benzothiophene ...

Indian Journal of Chemistry Vol. 47B, April 2008, pp. 586-591

QSAR studies on 4-thiazolidinones and 2-azetidinones bearing benzothiophene nucleus as potential anti-tubercular agents A S Narute , P B Khedekar & K P Bhusari* Sharad Pawar College of Pharmacy, Wanadongri, Hingna Road, Nagpur 441 110, India E-mail: [email protected] Received 31 March 2006; accepted (revised) 10 January 2008 Quantitative structure-activity relationships (QSAR) study on a series of (substituted 1, 2-dihydro)4–thiazolidinones and 2-azetidinones bearing benzothiophene nucleus with anti-tubercular activity has been carried out using a combination of various physicochemical descriptors. Several significant equations with good co-efficient of correlation (>0.860) have been obtained. The two models are selected using internal predictive power discerned by cross-validated coefficient q2. Both models highlight some common important feature, i.e., bulky substitution and the high nucleophilicity nature of the molecules, favorable for anti-tubercular activity. Keywords: Anti-tubercular activity, QSAR, benzothiaphene, anti-tubercular agents

Tuberculosis is a chronic grannulomatous disease1. It is estimated that today one-third to one-half of the world population is infected with tuberculosis leading to approximately 6% of all deaths worldwide2,3. The causative moiety of the disease is Mycobacterium tuberculosis4. Despite of the development of several types of synthetic anti-tubercular agents, the incidences of tuberculosis is still increasing in large parts of the world due to the development of resistance in Mycobacterium to the available drugs. Thus, there is an urgent need for novel anti-tubercular agents. With modes of action and chemical structures different from the currently used compounds, it is planned to study the quantitative structure activity relationships5,6 (QSAR), of some 4-thiazolidinones7-9 and 2azetidinones10-12, which have played an important role in medicinal chemistry. Moreover, they have been studied extensively because of their ready accessibility, diverse chemical reactivity and broad spectrum of biological activity. The focus of the present investigations is the QSAR analysis of thiazolidinones and azetidinones nucleus as potent anti-tubercular agents. Materials and Methods The Dataset and Parameters Quantitative structure activity relationship (QSAR) studies of anti-tubercular activity of newly reported

thiazolidinones and azetidinones derivatives against Mycobacterium tuberculosis reported by Joshi et al13 were performed using linear free energy relation of Hansch. Some of the compounds reported in the original paper were excluded in the present study because of their non-graded quantitative activity data or non-availability of parametric values. Antitubercular activity of remaining compounds are given in Table I. The biological activity values [BA] reported in the literature were converted to molar units and then further to -log scale and subsequently used as the response variable for the QSAR analysis. The molar anti-tubercular activities were then subjected to multiple regression analysis on different physicochemical parameters and indicator variables (QSAR). The relationships between the activities were also studied to explore the selectivity in terms of structural requirements. The congeneric series possesses one region of structural variation. Figure 1 shows effect of R substitution on 2-(substitutedbenzal- hydrazinocarbonyl)-3,5-chlorobenzo(b)thiophene, Figure 2 shows effect of R substitution on 2aryl-5H-3-(3′,5′-dichloro-2′-benzo(b)thiophenylamino)4-thiazolidin- ones and Figure 3 shows effect of R substitution on the 4-aryl-3-chloro-1-(3′,5′-dichloro2′-benzo(b)-thiophenylamino)-2- azetidinones. All the computations in the present study were performed on PIV workstation. The molecular structures of the training set were sketched using

NARUTE et al.: QSAR STUDIES ON 4-THIAZOLIDINONES AND 2-AZETIDINONES

Table I — Anti-tubercular activity data for 4-thiazolidinones and 2-azetidinones benzothiophene derivatives used in this study Compd 1a 1b 1c 1d 1e 1f 2a 2b 2c 2d 2e 2f 2g 2h 2i 3a 3b 3c 3d 3e 3f 3g 3h 3i

R 3-Br-C6H4 4-OCH3- C6H4 3-OCH3 4-OH- C6H3 4-N, N- (CH3) 2- C6H4 3-OC6H5- C6H4 4-S- CH3- C6H4 3-Br-C6H4 3-Cl- C6H4 3,4-(O CH3) 2- C6H3 4-OCH3- C6H4 3-OCH3 4-OH- C6H3 4-N, N- (CH3) 2- C6H4 3-OC6H5- C6H4 4-S- CH3- C6H4 3,4,5-(O CH3) 3- C6H2 3-Br-C6H4 3-Cl- C6H4 3,4-(O CH3) 2- C6H3 4-OCH3- C6H4 3-OCH3 4-OH- C6H3 4-N, N- (CH3) 2- C6H4 3-OC6H5- C6H4 4-S- CH3- C6H4 3,4,5-(O CH3) 3- C6H2

BA

-logBA

07 14 08 12 06 07 20 15 01 18 10 11 18 06 20 19 26 28 02 10 15 20 22 10

1.12 0.78 1.06 0.86 1.19 1.12 0.6 0.75 1.99 0.65 0.95 0.9 0.65 1.19 0.6 0.62 0.45 0.41 1.69 0.65 0.75 0.6 0.54 0.95

Chem Draw Ultra module of CS Chem Office 2001 molecular modeling software ver. 6.0, supplied by Cambridge Software Company14. The sketched structures were exported to Chem3D module in order to create its 3D model. Each model was “cleaned up” and energy minimization was performed using Allinger’s MM2 force field by fixing Root Mean Square Gradient (RMS) to 0.1 Kcal/molÅ. Further, geometry optimization was done using semiemperical AM1 (Austin Model) Hamiltonian method, closed shell restricted wave function available in the MOPAC module until the RMS value becomes smaller than 0.001 Kcal/molÅ. The low energy conformers obtained from the aforementioned procedure was used for the calculation of the descriptors. The descriptors include physicochemical, thermodynamic, electronic and spatial descriptors available in the ‘Analyze’ option of the Chem. 3D package (Table II). The descriptors calculated for the present study accounts for four important properties of the molecules: physicochemical, thermodynamic,

H

Cl

Cl

587

O

R N

N H S 1 a-f Figure 1 ⎯ 2-(Substituted-benzalhydrazinocarbonyl)-3,5chlorobenzo(b)thiophen

R

Cl

Cl

O

H

S H

N N H

S

H

O

2a-i Figure 2 ⎯ 2-Aryl-5H-3-(3′,5′-dichloro-2′benzo(b)thiophenylamino)-4-thiazolidinones

Cl

Cl

R O

S

H

Cl

N N H

H O

3 a-i Figure 3 ⎯ 4-Aryl-3-chloro-1-(3′,5′-dichloro-2′benzo(b)thiophenylamino)-2-azetidinones

electronic and steric, as they represent the possible molecular interactions between the receptor and thiadiazinoacridines. Multivariate Regression Analysis The regression analyses were carried out using SYSTAT15 version 10.2. The statistical quality of equation was judged by the parameters like correlation coefficient (R), standard deviation, standard error of estimation (SEE), variance ratio (f), at specified degree of freedom (df), and ‘t’ values of the regression constant (i.e., the constant term of the regression equation: regression coefficients and intercepts). The use of more than one variable in multivariate equation was justified by autocorrelation study. All the accepted equations have regression constant and f-ratio significant at 95% and 99% level, respectively, if not stated otherwise. Predicted Residual Analysis QSAR models can be cross-validated by predicted residual leave one out (LOO) analysis. Each compound of the list is deleted once from the data set

INDIAN J. CHEM., SEC B, APRIL 2008

588

Table II — Descriptors calculated for the QSAR study Sr NoDescriptor 1 2 3 4 5 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39

Heat of Formation (HF) Boiling Point (BP) Critical Pressure (CP) Critical Temperature (CT) Critical Volume (CV) Henry's Law Constant (HLC) Ideal Gas Thermal Capacity (IGTC) Log P Melting Point (MP) Molar Refractivity (MR) Standard Gibbs Free Energy (SGFE) Connolly Accessible Area (CAA) Connolly Molecular Area (CMA) Connolly Solvent–Excluded Volume (CSEV) Ovality (OVA) Principal Moment of Inertia – X (PMI–X) Principal Moment of Inertia – Y (PMI–Y) Principal Moment of Inertia – Z (PMI–Z) Dipole Moment (D) Dipole Moment –X Axis (DX) Dipole Moment –Y Axis (DY) Dipole Moment –Y Axis (DZ) Electronic Energy (EE) HOMO Energy (HOMO) LUMO Energy (LUMO) Repulsion Energy (RE) Bend Energy (Eb) Charge–Charge Energy (CCE) Charge–Dipole Energy (CDE) Dipole–Dipole Energy (DDE) Non–1, 4 VDW Energy (Ev) Stretch Energy (SE) Stretch–Bend Energy (SBE) Torsion Energy (Et) Total Energy (E) Van der Waals e 1,4 Energy (VDWE) VDW 1,4 Energy (VDWE) Partition coefficient

Table III — Calculated descriptor values for the given series of compounds

Type Thermodynamic Thermodynamic Thermodynamic Thermodynamic Thermodynamic Thermodynamic Thermodynamic Thermodynamic Thermodynamic Thermodynamic Thermodynamic Steric Steric Steric Steric Steric Steric Steric Electronic Electronic Electronic Electronic Electronic Electronic Electronic Electronic Thermodynamic Thermodynamic Thermodynamic Thermodynamic Thermodynamic Thermodynamic Thermodynamic Thermodynamic Thermodynamic Thermodynamic Thermodynamic Thermodynamic

and corresponding regression equation is found out to calculate predicted activity value and predicted residual (press) of deleted compound. The PRESS (predicted residual sum of squares) statistics provides the relations between the observed activity and calculated value (according to PRESS equation). The

Compd

PMI-X

D

HOMO

Et

1a 1b 1c 1d 1e 1f 2a 2b 2c 2d 2e 2f 2g 2h 2i 3a 3b 3c 3d 3e 3f 3g 3h 3i

2730.62 2160.68 2288.34 2343.45 3051.31 2252.2 2699.54 2527.28 3154.93 2425.3 2947.99 2754.77 2353.77 2945.17 3297.32 2998.69 1826.95 3052.83 2659.45 2281 3068.56 3224.63 2523.64 3644.38

2.4079 3.5584 5.1355 6.0749 3.2713 3.7861 3.7577 3.414 3.8617 4.4814 3.5869 4.5916 3.0341 2.6997 4.3984 4.1765 4.1287 4.3274 4.5517 4.6457 7.1458 5.5351 6.2512 4.2893

-8.7059 -8.6642 -8.7057 -8.26 -8.7667 -8.2024 -8.8652 -8.6927 -8.7583 -8.8369 -8.6887 -8.3607 -8.7454 -8.1723 -8.8235 -8.8799 -8.9326 -8.8261 -8.8324 -8.667 -8.4923 -8.9489 -8.6158 -8.7998

-5.2672 -5.3606 -6.1963 -7.5633 -11.004 -7.967 1.76074 2.42927 4.29722 1.78119 0.47622 2.07256 -2.6077 -1.7454 3.87878 5.07532 5.83832 4.15428 5.93828 3.66351 2.65777 -1.2288 3.63627 6.50846

stability and predictive capacity of the equation were cross validated from PRESS statistics obtained by running VALSTAT16 programs using “leave-one-out” technique. Results and Discussion Biological activity data and various physicochemical parameters were taken as dependent and independent variables, respectively, and correlations were established using sequential multiple regression analysis. The descriptors selected for modeling antitubercular activity of thiazolidinones and azetidinones derivatives are summarized in Table III. The quarter parametric models were obtained and these models are significant for anti-tubercular activity. Model I -logBA = 4.48858(± 2.70123) + 0.00018139(± 0.000147211) PMI-X - 0.0606539(± 0.0559619) D +

NARUTE et al.: QSAR STUDIES ON 4-THIAZOLIDINONES AND 2-AZETIDINONES

0.453772(±0.305151)HOMO-0.0246623(± 0.014462) Et n =22, R=0.860104, Variance=0.0183792, SD =0.13557, F=12.0823 Model II -logBA = 4.45369(± 2.67143) + 0.000217417(± 0.000135005) PMX- 0.0646233(± 0.0528988) D + 0.458231(±0.301616)HOME- 0.0232831(± 0.014025) Et n=24, R =0.854129, variance=0.0183422, SD =0.135433, F=12.8125 The study of model I and model II revels those thermodynamic parameters like torsion energy (Et), steric parameters like principal moment of inertia Xaxis (PMI-X) and electronic parameters like dipole moment (D) and highest occupied molecular orbit (HOMO) are associated with anti-tumor activity. In model I, dipole moment, an electronic parameter and is important in case when dipole– dipole interactions are involved in ligand–receptor interactions and torsion energy (Et) is the thermodynamic parameter, which represents the energy associated with deforming torsion angles in the molecules from their ideal values. The negative coefficients of descriptors suggest presence of conjugation and bulky substituents tolerable for activity, whereas principal moment of inertia, X-axis is a spatial descriptor, which explains the significance of orientation and conformation rigidity of the molecule. The positive coefficient of these descriptor suggest the presence of bulky substituents oriented towards X-axis of the molecules will give better activity and highest occupied molecular orbital. HOMO is an electronic parameter and is the highest energy level in the molecule that contains electrons. It is crucially important in governing the molecular reactivity and properties. When a molecule acts as an electron pair donor in bond formation, the electrons are supplied form the molecule’s HOMO. HOMO descriptor denotes nucleophilicity of the molecule and this term was correlated positively. The model suggests that PMIX and HOMO is of significance having high value of t-test indicating statistical significance of calculated regression coefficient. To confirm this result the value of -logBA was estimated using LOO and correlated with observed value of -logBA. The value of bootstrapping r2, chance and q2 in randomized biological activity indicates statistical significance of model as follows.

589

Bootstrapping r2 = 0.813112, q2 = 0.571626, Spress = 0.173942, SDEP = 0.152903 In model II principal moment of inertia and highest occupied molecular orbital positively contribute to biological activity whereas dipole moment and torsion energy negatively contribute to biological activity. The model suggests that principal moment of inertia and highest occupied molecular orbital is of significance having high value if t-test indicating statistically significant calculated regression coefficient. Leave one out cross validation method was used for predictivity of model II. The value of bootstrapping r2, chance and q2 in randomized biological activity indicates the statistical significance of the model as follows. Bootstrapping r2 = 0.777145, q2 = 0.586244, Spress = 0.167511, SDEP = 0.149044 The correlation matrix shows model II to be more significant than model I. In model I, all independent parameters (PMI-X, D, Et and HOMO) have poor (independent) correlation with each other as expected in QSAR analyses but in model-II independent parameters (PMI-X, D, Et and HOMO) have dependent correlation. The correlation matrix and predicted activity data for model I and model II are shown in Table IV, V and Table VI, VII, respectively. Figure 4 and Figure 5 show a plot of observed vs predicted activities of compounds of model I and model II, respectively. The comparison of model I and model II, the model II was more significant than model I, having good correlation coefficient (R), crossvalidated (q2) value (reflects predictive power of model) bootstrapping (r2) value (reflect accuracy of the model), and independent correlation between Table IV — Correlation matrix for parameters in model I Parameters

PMI-X

D

HOMO

Et

PMI-X D HOMO Et

1.000000 0.055336 0.170548 0.256476

1.000000 0.077995 0.222781

1.000000 0.402842

1.000000

Table V — Correlation matrix for parameters in Model II Parameters

PMI-X

D

HOM0

Et

PMI-X D HOMO Et

1.000000 0.014297 0.208958 0.319568

1.000000 0.090494 0.247327

1.000000 0.405912

1.000000

INDIAN J. CHEM., SEC B, APRIL 2008

590

Table VI — Predicted activity data of model I Predicted -logBA

1.12 1a 0.78 1b 1.06 1c 0.86 1d 1.19 1e 1.12 1f 0.6 2a 0.75 2b 0.65 2d 0.95 2e 0.9 2f 0.65 2g 1.19 2h 0.6 2i 0.62 3a 0.45 3b 0.41 3d 0.65 3e 0.75 3f 0.6 3g 0.54 3h 0.95 3i Compd 2c and 3c are outlier

0.991225 0.881516 0.717722 1.05483 1.09521 1.15091 0.692616 0.733646 0.597315 0.842528 0.856629 0.855464 1.19709 0.74133 0.625229 0.31487 0.702978 0.588495 0.658391 0.754452 0.575164 0.643861

1.01723 0.865298 0.794575 0.983557 1.13694 1.14194 0.684136 0.735515 0.602833 0.851327 0.864789 0.827379 1.19375 0.720402 0.624585 0.372207 0.67239 0.597353 0.692664 0.707332 0.567922 0.735868

Table VII — Predicted activity data of Model II Compd 1a 1b 1c 1d 1e 1f 2a 2b 2c 2d 2e 2f 2g 2h 2i 3a 3b 3c 3d 3e 3f 3g 3h 3i

Observed -log BA 1.12 0.78 1.06 0.86 1.19 1.12 0.6 0.75 1.99 0.65 0.95 0.9 0.65 1.19 0.6 0.62 0.45 0.41 1.69 0.65 0.75 0.6 0.54 0.95

Predicted -logBA 1.00262 0.860242 0.698692 1.01338 1.11086 1.12779 0.703618 0.741783 0.595105 0.864143 0.871481 0.84961 1.23674 0.772968 0.651867 0.289744 0.723234 0.583873 0.684702 0.776434 0.57102 0.735432 0.57102 0.735432

1.4

Calculated -logBA

Calculated -logBA 1.02507 0.848092 0.77437 0.961711 1.14472 1.1255s9 0.694477 0.742706 0.870313 0.876491 0.822645 1.2154 0.752844 0.648548 0.354963 0.696674 0.5926 0.705752 0.72504 0.565729 0.78498 0.565729 0.78498 0.870313

1.2 Predicted activity

Observedlog BA

1 0.8 0.6 0.4 0.2 0 0

0.5

1

1.5

Observed activity

Figure 4 — Graph between observed activity and predicted activity of model I 1.4 Predicted activity

Compd

1.2 1 0.8 0.6 0.4 0.2 0 0

0.5

1

1.5

Observed activity

Figure 5 — Graph between observed activity and predicted activity of model II

parameters as expected in QSAR analyses. These results show that such models can be helpful for theoretical prediction of anti-tubercular activity of new molecules. Conclusion QSAR analysis was performed on a series of antitubercular activity of thiazolidinones and azetidinones derivatives using molecular modeling program Chemoffice 2001. QSAR models were proposed for anti-tubercular activity of the thiazolidinones and azetidinones using descriptors employing sequential multiple regression analysis method. The predictive power of each model was estimated with bootstrapping r2 method and leave one out cross validation method. It was observed from the selected models that biological activity of thiazolidinones and azetidinones derivatives is governed by thermodynamic and steric properties of the molecules. The models also provide valuable

NARUTE et al.: QSAR STUDIES ON 4-THIAZOLIDINONES AND 2-AZETIDINONES

insight into the mechanism of action of these compounds. The result of the study suggests involvement of partition coefficient in the mechanism of anti-tubercular action of thiazolidinones and azetidinones. The study will be helpful in the design of better anti-tubercular analogs of thiazolidinones and azetidinones derivatives for anti-tubercular activity. Acknowledgements One of the authors, Ashok Narute, is grateful to the All India Council for Technical Education (AICTE) for providing fellowship. The authors wish to especially thank Dr. P. Trivedi and Director, Department of Pharmacy, Shri G.S. Institute of Technology and Sciences for providing software for study and Principal, Sharad Pawar College of Pharmacy, Nagpur for providing the necessary facilities for undertaking this research work. References 1 Tripathi K D, Essentials of Medical Pharmacology, 5th Edn (Jaypees Brothers Medical Publishers Pvt Ltd, New Delhi), 2003, 689. 2 Marwick C, JAMA, 267, 1992, 174.

591

3 Daniel Y M, Tuberculosis in Harrison’s Principles of Internal Medicine, 12th Edn, edited by Jean Wilson D (McGraw-Hill, New York), 1991. 4 Williams D A & Lemke T L, Foye’s Principle of Medicinal Chemistry, 5th Edn (Lippincott Wiliams and Wilkins, New York), 2002, 905. 5 Hansch C, Comprehensive Medicinal Chemistry (Paragon Press, New York), 4, 1990, 579. 6 Smith H J & Williams H, Introduction to Principle of Drug Design (John Wright and Sons Ltd, Bristol), 1983, 216. 7 Sharma R C & Kumar D, J Inst Chem Soc, 77, 2000, 492. 8 Joshi H, Upadhay P S & Baxi A J, Indian J Chem, 39B, 2000, 967. 9 Ingle S, Sawale A R, Ingle R D & Mane R A, Indian J Chem, 40B, 2001, 124. 10 Kagathara P, Upadhay T, Doshi R & Parekh H H, Indian J Het Chem, 10, 2000, 9. 11 Matsui N, Jpn Kokai Tokyo JP, 07, 2000, 652; Chem Abstr, 132, 2000, 64109u. 12 Desai K R, Asian J Chem, 132, 2000, 279145. 13 Joshi H S, Thaker K M & Kachhadia V V, Indian J Chem, 42B, 2003, 1544. 14 CS Chem Office, version 6.0, Cambridge Soft Corporation, software publisher Association, 1730 M Street, NW, Suite 700, Washington DC, 20036 (202), 452-1600, USA. 15 SYSTAT 10.2 version supplied by Systat Software Inc. 16 Gupta A K, Babu M A & Kaskhedikar S G, VALSTAT: Validation Program for Quantitative Structure Activity Relationship Studies, Indian J Pharm Sci, 66, 2004, 396.

Suggest Documents